What’s next for AI in the finance function?

By Vin Kumar, below left, and John O’Mahony, right, The Hackett Group
In 2023, artificial intelligence (AI) was a topic that floated around the office, but not something that landed on the desk of many chief financial officers. That’s about to change. We believe 2024 will be the year that generative AI (Gen AI) becomes an integral part of every sophisticated finance department. The addition of Gen AI capabilities to familiar applications, the power to use AI to develop basic software without coding, and the ability to collaborate with AI programmes trained on tax, audit, and other specialised bodies of regulatory knowledge will lead to dramatic changes in how finance operates.
We expect to see three varieties of Gen AI in the market next year – each of which promises new ways to make finance operate more efficiently and effectively:
Embedded Gen AI  
Developers of most of the software you use every day are now working on adding AI functionality to their existing system. This won’t replace people, but it will make some individuals more productive and effective.
Among other things, many common tools will now include Microsoft’s Gen AI-powered Copilot feature or something analogous to it. These will give you the ability to:

Order a document reviewed and summarised. “ Summarise5 the key points of this contract and tell me about its liability clauses.”


Create a computer program on the fly that can manipulate data from multiple sources and put it in a report. “ I need a Visual Basic program that will take this data from these business units and compare it to our overall performance in the following format.”


Take care of your basic correspondence. “Write an email thanking Kevin for his 15 years of service. Ask him if he’s free for lunch next Thursday.”

The good news here is that the developers will add these features without any need for planning or investment on your part. Your team will just have to learn how to use them when they are released.
Native Gen AI  
Working directly with a Gen AI model, you will be able to develop programmes that will speed up your work in a variety of ways such as:

Generate the comments and notes in your regulatory compliance statements.


Create monthly business unit reports, freeing your analysts to spend more time on in-person conversations with the business unit.


Update your rules engines for your enterprise resource planning systems easily and on an ongoing basis. Instead of having to wait for annual updates, you will be able to add new exceptions on the fly without having to rely on an information technology person.

Domain-specific Gen AI  
Beginning in the second half of the year, we expect to see specialised large language models reach the market. Instead of being trained on a wide variety of material, these models will contain comprehensive knowledge on a particular specialised subject such as internal auditing, tax, or environmental, sustainability, and governance reporting. They will then be able to apply that knowledge to create reports that will do things such as:

Speed up filings.


Generate reports with lower rates of errors (“hallucinations”) than other large language models thanks to their narrow professional focus.


Keep your data secure. The providers (who are mostly well-known accounting and audit firms) are well aware of the need for cyber safeguards.

On your mark  
Of course, to take full advantage of these three kinds of AI, your department will first need to apply a certain amount of the organic variety. To be prepared for this wave of opportunity, you should:

Identify people on your team who might be good at experimenting with these new tools. Sign them up for one of the free courses on Gen AI basics offered by Google, Amazon, Microsoft and the other big technology companies. You want to make sure they are ready to start using this next generation of Gen AI tools when they are rolled out.


Reach out to your vendors and ask them what they are doing with AI. If you work with business process outsourcers, for instance, sound them out about their Gen AI plans. Impress upon them that you need to know their plans so you don’t duplicate efforts.


Focus on identifying what you will need to do to create good proofs of concept. Do you have the right data, right talent and right technology? Do you have a good sense of how much it is going to cost, or how much money or effort it will save? Keep in mind that unlike past technologies, Gen AI can only offer answers that are probabilistic – not quantitative. This opens a wide range of new opportunities, but also creates new sets of risks that are still not entirely well understood.

Once you have worked these preparations, set an aggressive timeline for your proof of concepts and those of your partners. We expect that by Q4 2024, leading finance functions will have a much more concrete sense of their long-term plans for Gen AI.
With so many new capabilities on the way, it’s difficult to know in advance which tools will matter most. It’s safe to assume, however, that chance is likely to favour the prepared team.
Vin Kumar is The Hackett Group’s AI and Global Digital Operations Practice Leader. John O’ Mahony is The Hackett Group’s Head of Finance and Business Transformation
 
The post What’s next for AI in the finance function? appeared first on SME Magazine .